Member Junction
    Preparing search index...

    Interface TranscriptSegmentationOptions

    Options specific to TranscriptSegmenter.

    interface TranscriptSegmentationOptions {
        BoundaryGapMs?: number;
        DefaultModality?: "audio" | "video";
        EmitSubChapters?: boolean;
        MaxChapterMs?: number;
        MaxSegmentTokens?: number;
        MinSegmentTokens?: number;
        OverlapTokens?: number;
    }

    Hierarchy (View Summary)

    Index

    Properties

    BoundaryGapMs? DefaultModality? EmitSubChapters? MaxChapterMs? MaxSegmentTokens? MinSegmentTokens? OverlapTokens?

    Properties

    BoundaryGapMs?: number

    A silence gap at least this long is treated as a chapter boundary — the cheapest reliable topic-shift signal in a recording. Default: 4000.

    DefaultModality?: "audio" | "video"

    Modality to stamp on emitted segments when it can't be inferred. Default: 'audio'.

    EmitSubChapters?: boolean

    Also emit one child segment per speaker turn inside each chapter. Doubles embedding cost, so it is opt-in; enable for panel/interview content where per-speaker retrieval matters. Default: false.

    MaxChapterMs?: number

    Maximum wall-clock length of one chapter, in ms. Default: 300000 (5 minutes).

    MaxSegmentTokens?: number

    Hard ceiling on tokens per text segment. The base class splits any oversized segment via TextChunker so no segmenter can exceed it. Default: 512.

    MinSegmentTokens?: number

    Segments whose text estimates below this many tokens are merged forward into the next segment, preventing a spray of near-empty vectors. Default: 0 (off).

    OverlapTokens?: number

    Overlap tokens applied when an oversized segment must be split. Default: 10% of max.

    Settings

    Member Visibility

    On This Page

    Properties